Dasatinib response in acute myeloid leukemia is correlated with FLT3/ITD, PTPN11 mutations and a unique gene expression signature

Abstract

Novel targeted therapies demonstrate improved survival in specific subgroups (defined by genetic variants) of acute myeloid leukemia (AML) patients, validating the paradigm of molecularly targeted therapy. However, identifying correlations between AML molecular attributes and effective therapies is challenging. Recent advances in high-throughput in vitro drug sensitivity screening applied to primary AML blasts were used to uncover such correlations; however, these methods cannot predict the response of leukemic stem cells (LSCs). Our study aimed to predict in vitro response to targeted therapies, based on molecular markers, with subsequent validation in LSCs. We performed ex vivo sensitivity screening to 46 drugs on 29 primary AML samples at diagnosis or relapse. Using unsupervised hierarchical clustering analysis we identified group with sensitivity to several tyrosine kinase inhibitors (TKIs), including the multi-TKI, dasatinib, and searched for correlations between dasatinib response, exome sequencing and gene expression from our dataset and from the Beat AML dataset. Unsupervised hierarchical clustering analysis of gene expression resulted in clustering of dasatinib responders and non-responders. In vitro response to dasatinib could be predicted based on gene expression (AUC=0.78). Furthermore, mutations in FLT3/ITD and PTPN11 were enriched in the dasatinib sensitive samples as opposed to mutations in TP53 which were enriched in resistant samples. Based on these results, we selected FLT3/ITD AML samples and injected them to NSG-SGM3 mice. Our results demonstrate that in a subgroup of FLT3/ITD AML (4 out of 9) dasatinib significantly inhibits LSC engraftment. In summary we show that dasatinib has an anti-leukemic effect both on bulk blasts and, more importantly, LSCs from a subset of AML patients that can be identified based on mutational and expression profiles. Our data provide a rational basis for clinical trials of dasatinib in a molecularly selected subset of AML patients.

Figures

Figure 1.

In vitro drug resistance. (A) In vitro drug resistance to most drugs is acquired after relapse. An average drug response of seven couples of primary acute myeloid leukemia (AML) samples, at diagnosis (DX) and relapse (REL) is shown, comparing the half maximal inhibitory concentration (IC50) of 41 drugs. Each dot represents the response to a specific drug, calculated by the median IC50 ratio of diagnosis vs. relapse. (B) Drug sensitivity and resistance testing (DSRT) of primary AML cells. Hierarchical clustering using Pearson dissimilarity and complete linkage was performed. Data are the log2 IC50 +0.001, standardized for each compound by reducing the mean. DSRT for 41 drugs of clinical and preclinical use in AML shows two clustered groups of patients. The age and gender of the patients and the origin of the patients’ diagnostic or relapse sample, the name of the drugs, the class of the drug, karyotype and mutation status for FLT3/ITD and NPM1 are shown.

Figure 2.

Transcriptome profiling of dasatinib responder samples. Gene expression analysis of whole transcriptome mRNA sequencing comparing dasatinib sensitive to non-sensitive samples of acute myeloid leukemia (AML). (A) Differentially expressed genes between "responders", and "non-responders". (B) In order to extend the tested samples we also applied the same analysis to AML patients’ samples from the Beat AML dataset.

Figure 3.

Analysis of differentially expressed genes between dasatinib responders and non-responders from the INCPM and Beat AML datasets. (A) Intersection of upregulated (upper panel) and downregulated (lower panel) differentially expressed genes from the INCPM and Beat AML cohorts. (B) Pathway analysis of intersecting upregulated genes in both cohorts identified significant enrichment of genes that are co-expressed with several dasatinib targets. Dasatinib targets (CSF1R, SRC, BLK) are shown in red and the asterisk marks significant enrichment (false discovery rate<0.05).

Figure 4.

Prediction model to identify dasatinib responders based on gene expression. (A, B) The k-nearest neighbor (KNN) algorithm with cosine similarity was used to predict the response to dasatinib. The greatest accuracy was achieved with 10- fold cross-validation applied to the differentially downregulated genes from the Beat AML dataset. Accordingly, a sensitivity of 0.85 and a specificity of 0.7 were achieved. (C) Validation of the KNN cosine prediction model on INCPM data.

Figure 5.

Correlation between FLT3/ITD, PTPN11 and TP53 mutations in acute myeloid leukemia samples and response to dasatinib in vitro in the Beat AML study. (A-C) Median dasatinib half maximal inhibitory concentration (IC50) values in cases with mutated FLT3/ITD (A), PTPN11 (B) or TP53 (C) and wild-type (WT) samples. Differences between groups were validated by the Wilcoxon rank-sum test.

Figure 6.

Dasatinib delays development of human FLT3/ITD-mutated acute myeloid leukemia in transplanted mice. SGM3 mice were treated with dasatinib for 3 weeks after transplantation of human cells from nine patients with FLT3/ITD acute myeloid leukemia (AML). The percentage of human CD45+ (hCD45) cells in engrafted murine bone marrow with/without dasatinib treatment is shown after staining for CD45, CD34, CD33, CD38, CD15, CD3 and CD19 to determine myeloid lineage cells and analyzed by fluorescence- activated cell sorting immunostaining. Wilcoxon rank sum test, *P<0.05; **P<0.005; ***P<0.0005. The mutation status of the AML samples was: 140005: FLT3/ITD+, FLT3/TKD-, NPM1+; 40094: NPM1+, FLT3/ITD+; 150809: NPM1-, FLT3/ITD+; 40034: NPM1-, FLT3/ITD+, 150279: FLT3/ITD+, FLT3/TKD-, NPM1+; 130695: NPM1+, FLT3/ITD+, FLT3/TKD-; 160436: NPM1+, FLT3/ITD+; 160406: FLT3/ITD+, NPM1+; 130607: FLT3/ITD+, NPM1+.

Graphical Abstract